Good suggestion. At least if I understand what you mean

Something along the lines of

thenChar c f = f >>= (\ r -> char c >> return r)

with a specialised one for spaces maybe

thenSpace = thenChar ' '

I suppose liftM and friends can be employed to remove the function calling in the creation of Address, Device and MemRegion. I’ll try to venture into Parsec territory once again soon and report on my findings.

Here’s an idea that for getting more functional/applicative formulations. Replace all of the explicitly sequential (”do”) parsec code with liftM, liftM2, …. From a quick read-through, I think you can do it. For parseRegion, you could use an auxiliary function that discards a following space, e.g. “thenSpace (many1 digit)”.

Also, play with factoring out some of the repeated patterns in parseAddress, parsePerms and parseDevice.

The more I play with refactoring in my code, the more elegant it gets and the more insight I get. Have fun!